Software for hepatic vessel classification: feasibility study for virtual surgery.

Purpose: The detection and classification of hepatic vessels in diagnostic images are essential for hepatic pre-surgery planning. Our team has developed a tool for classification, analysis, and 3D reconstruction of the hepatic and portal systems. Methods: Our software first extracts a graphic representation of a set of connected voxels, representing both systems. It then calculates two binary volumes representing the main part of the two venous systems. Finally, it combines these results to obtain the correct vessel classification. Results: Segmentation steps are semi-automatic and require about 40 min to complete. Schematization and classification steps are automatic and require about 17 min for results. Conclusion The software provides a correct and detailed reconstruction even where pathologies have caused morphological and geometrical variations in the vessels. The time required for the entire procedure is compatible with clinical requirements, providing an efficient tool for diagnosis and surgical planning.